Comment 11 for bug 1730717

@Laney, thanks for testing the mainline kernel. It's promising that a fix might be in that kernel. The time consuming part will be identifying what commit in that kernel is the actual fix. We could perform a "Reverse" kernel bisect, which would required testing 12 or so test kernels. However, it sounds like to set up a reproducer is time consuming as well.

The easiest thing to try next would be to test the latest upstream 4.13 stable kernels. It's possible the fix that is in 4.14-rc8 was also cc'd to upstream stable and made it's way into 4.13 thorough the normal stable update process.

If possible to test, the latest 4.13 upstream kernel is 4.13.12 and is available here:
http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.13.12/